[
https://issues.apache.org/jira/browse/LUCENE-2184?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12795080#action_12795080
]
Chris Male commented on LUCENE-2184:
------------------------------------
I have also seen this bug come up. It seems that there are only 15 possible
tiers, as if the best fit tier is greater than 15, then it is scaled back to
15. So one plausible solution that doesn't require any changes to the API, is
to simply require all documents have all 15 levels indexed. The other solution
is to take into account the range of indexed levels when calculating the the
best fit tier, but this would require an API change, although it could be made
backwards compatible quite easily.
Any thoughts on the preferred approach?
> CartesianPolyFilterBuilder doesn't properly account for which tiers actually
> exist in the index
> ------------------------------------------------------------------------------------------------
>
> Key: LUCENE-2184
> URL: https://issues.apache.org/jira/browse/LUCENE-2184
> Project: Lucene - Java
> Issue Type: Bug
> Components: contrib/spatial
> Affects Versions: 2.9, 2.9.1, 3.0
> Reporter: Grant Ingersoll
>
> In the CartesianShapeFilterBuilder, there is logic that determines the "best
> fit" tier to create the Filter against. However, it does not account for
> which fields actually exist in the index when doing so. For instance, if you
> index tiers 1 through 10, but then choose a very small radius to restrict the
> space to, it will likely choose a tier like 15 or 16, which of course does
> not exist.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: java-dev-unsubscr...@lucene.apache.org
For additional commands, e-mail: java-dev-h...@lucene.apache.org
